| T O P I C R E V I E W |
| index0192 |
Posted - 12/16/2022 : 01:46:25 AM
Dear all,
I have some question about the cumulative fit peak value when I use the multiple peak fit. I start with my data (x,y) likes this
(there is more data than what you can see here, but I just take a picture to show you how my data look like)
after I draw the plot, I use the following steps:
the Analysis -> multiple peak fit -> then I use three gaussian to fit the data
then the data and fit look like this
Now, after fitting, I read the fit detail in the new page of the data book, which looks like this
and my questions come
(1) why the values in the A(X1) column is not my original x-data, although the F(X2) column is indeed my original x-data?
(2) why the values in E(Y1) column, i.e the Cumulative Fit Peak, is not equal to the values of B(Y1) + C(Y1) + D(Y1), i.e peak1 + peak2 + peak3?
Could someone tell me how the origin calculate this column?
Thank you!
|
| 3 L A T E S T R E P L I E S (Newest First) |
| Echo_Chu |
Posted - 12/19/2022 : 06:29:28 AM
Hi,
You can also control the X range of fitted curve with Multiple Peak Fit tool.
Please click the Open NLFit button when you use Multiple Peak Fit, then in the opened dialog,
User's input data are quite different. so by default we don't use user's data as we don't know whether it will introduce problem. we just want to be safe for default settings.
However, there are options that you can get what you want.
In regard to the peak values. we have to add baseline because the fitted peak curves will be off the original curve without baseline.
Echo
OriginLab Technical Service
|
| index0192 |
Posted - 12/16/2022 : 11:15:11 AM
Dear Echo,
Thanks for answering me,
It helps a lot!
So back to question-1.
Yes I try to use "Peak Analyzer" rahter than "Multiple Peak Fit" as you suggested. Here are the steps I did (only mention the changes I made, ignore the unchanged options compared the defaut setting)
click my plot -> Analysis -> Peaks and Baseline
-> Peak Analyzer -> Open dialog
So I saw the dialog similar to your picture
Then I choose "Fit Peak(Pro)" in Goal
-> Next -> Next -> Next
-> disable the "Enable Auto Find", then "add" three peaks manually
-> choose "Positive" in Direction in Peak Finding Setting -> Next
-> choose "Same as Input Data" in X Data Type of Fitted Curves Plot of Configure Graph (as you said)
-> disable the "Use Separate X data for individual Peak" of Fitted Curves Plot
-> Fit Control --> Fit until converged -> OK -> Finish
and now I can see the fitting result like this
and when I look the detail page(name="FitPeakCurve1") in workbook as
I can see that the column A(X1) is the same of the column G(X2)
as you promised. Thanks!
But here comes some new questions:
Question-1.1
Why origin want to create another less bins X-data when I use "Multiple Peak Fit"? (what the motivation to have this function in origin)
Question-1.2
How does Origin decide the number of bins when it create this less bins X-data?
Question-1.3
For the purpose of using the original X-data in fitting detail page (I mean DON'T create another less bins X-data), is "Peak Analyzer" the only way to do that? Can I achieve the same thing by using "Multiple Peak Fit"? (is there some setting to do that in "Multiple Peak Fit")
Question-1.4
If I use "Peak Analyzer" rather than "Multiple Peak Fit", can I get the plot looks like this?
since I prefer this plot style more than the one I see from the "Peak Analyzer", I hope there is some setting can help.
Question-1.5
For the most important point, if it is the case that A(X1) column is not my original x-data (when I use "Multiple Peak Fit") as below,
how can I re-calculate the Reduced chi-square (I just want to calculate it by my hand and compare it with the value given by Origin)?
Since the procedure needs Y_observed_ith - Y_predicted_ith, however, these two values (Y_observed_ith and Y_predicted_ith) seems to be associated with different X-data (since E(Y1) the "Cumulative Fit Peak", which is Y_predicted I think, is associated with the new X-data, not the original X-data, but the only Y_observed we have is associated with the original X-data).
continuing the Question-2:
Ok, I get your point. So I understand that when I watch three peaks' values, the same baseline value are already included in them (the three peaks they share the same baseline value). Is it correct?
Therefore, If I directly sum peak1, peak2, peak3, since the baseline value is added three times, I can artificially subtract the baseline value twice, then the final result SHOULD equal to the value of "Cumulative Fit Peak". I have tried this and they are Equal! I think this also confirm my above statement, right?
(about the three peak share the same baseline, the way that Origin showing is really easy to make people confused, why you Origin want to do this? Have you ever clarified this somewhere in Origin official website or tutorial? I guess there are some people need it just like me)
Thanks! |
| Echo_Chu |
Posted - 12/16/2022 : 05:39:58 AM
Hi,
1. You can set the range of fitted curve to be the same as input data. then you will get A(X1) column the same as your original x-data. (See settings in image below)
2. The values of B, C, D column sum baseline in. If you add B,C,D directly, it is to add baseline three times. It won't equal to value of the cumulative peak curve.
Echo
OriginLab Techinical Service |
|
|
